Economic analysis of law in North America, Europe and Israel
Transcript of Economic analysis of law in North America, Europe and Israel
Economic analysis of standard form contracts:the monopoly case
Oren Gazal-Ayal
Published online: 22 September 2007
� Springer Science+Business Media, LLC 2007
Abstract The terms of standard form contracts are rarely known to consumers.
Still, it is often argued that few consumers who read and understand the contract can
assure that it does not include sub-optimal terms. According to this argument, if the
proportion of informed consumers is sufficiently high, they can secure an optimal
set of contract terms to the benefit of all other consumers. This paper shows that
when suppliers can adjust the content of the form contract, the few reading con-
sumers cannot correct the market failure. In fact, unless all consumers read and
understand the form contract, a monopoly is always encouraged to offer sub-optimal
terms, i.e., terms that benefit her but at a higher cost to the consumers.
Keywords Standard form contracts
JEL Classification K12 � D82
1 Introduction
The decades-long legal debate over the qualities and drawbacks of standard form
contracts is far from conclusion.1 At first, commentators concentrated on the
unequal bargaining power and the inability of consumers to bargain over the terms.2
O. Gazal-Ayal (&)
Faculty of Law, University of Haifa, Haifa, Israel
e-mail: [email protected]
1 The most cited early writings discussing form contracts are Karl N. Llewellyn, Book Review, 52 Harv.
L. Rev. 700 (1939), Friedrich Kessler, Contracts of Adhesion-Some Thoughts about Freedom of Contract,43 Colum. L. Rev. 629 (1943). However, the issue of standardizing relationships was addressed
previously by Nathan Isaacs, The Standardizing of Contracts 27 Yale L.J. 34 (1917).2 See Llewellyn, supra note 1, at 701, Kesler, supra note 1, at 632.
123
Eur J Law Econ (2007) 24:119–136
DOI 10.1007/s10657-007-9030-x
However, it was later shown that, other things being equal, sellers prefer to extract
their bargaining advantages by increasing the price rather than reducing the value of
the contract to the consumers (see Posner 1992). Thus, even a monopoly would
allocate the rights and risks in the contract efficiently. As a result, the emphasis has
shifted to the information problem.
Consumers often sign a contract without reading it, or without understanding the
implication of each term. Goldberg (1974) argues that since contract terms are often
unnoticeable, sellers are encouraged to offer low quality–low price contracts.
Moreover, cognitive deficiencies prevent consumers from fully internalizing the costs
of many terms, even if they know and understand them (Cruz and Hinck 1996; Hillman
and Rachlinski 2002; Korobkin 2003). The information asymmetry pushes sellers to
offer inefficient and disadvantaging terms to consumers.
Nevertheless, the common belief in the current literature is that this is usually not
an acute problem. Schwartz and Wilde (1979a), Trebilcock and Dewees (1981)
Gillette (2004) and many others argued that the few reading consumers could often
correct the market failure. They claimed that when the percentage of informed
consumers meet a certain threshold, the seller will behave as if all consumers are
informed. They even estimated that in some cases, it is enough that 33% of the
consumers, or even only 10%, read the contract, in order to correct the information
deficit (Trebilcock and Dewees 1981; Schwartz and Wilde 1979a). They claim that
even if the seller is a monopolist, the informed consumers can lead her to produce
efficient contract terms and exert her monopoly power on the price (see Trebilcock
and Dewees 1981; Schwartz 1977; Kornhauser 1976).3 This analysis dominates the
current legal thought.
It is true that a few scholars, including Korobkin (2003) and Cruz and Hinck
(1996) argue that sufficient number of fully informed and rational consumers is not
likely to be found. Yet even they do not argue against the basic claim of Schwartz
and Wilde (1979a) that the market might behave as if all consumers were informed
even if many of them were not.
In this paper I argue that unless all consumers are informed, sellers would offer
sub-optimal terms in the contract. I argue that the dominant view, according to
which an informed minority could assure efficient contracts, is based on models that
use unrealistic assumptions. This view relies mainly on the assumption that all
consumers have the same reservation price for the contract. When a more realistic
assumption of a down-sloping demand-curve is introduced, the result differs
substantially. As long as some of the consumers are uninformed, a monopoly will
offer inefficient terms in her standard form contract, i.e., terms that reduce the value
3 For a different view, see Cooter and Ulen (1997). They claim, following Kessler (1943), that standard
form contracts can be a tool to help a cartel to control its members. However, as long as the form contract
is not designated for cartelizing the market, it is, in their view, a tool that promotes efficiency. For
different uses of this argument, see also Kornhauser (1976). For other strategic uses of standard form
contracts, see Ahdieh (2005) and Gilo and Porat (2006). For a brief review of the historical development
of the normative literature on the subject, see De Geest (1994). See also De Geest (2002)
120 Eur J Law Econ (2007) 24:119–136
123
of the contract to consumers more than it reduces the costs to the seller.4 As I show,
the seller can adjust the content of each term to the proportion of consumers who are
informed of it. Hence, a higher proportion of informed consumers can reduce the
inefficiency, but as long as not all consumers are informed, the readers cannot fully
correct it. Terms that are often known to most consumers would be less inefficient
than terms that are known to very few. Thus, the two main points of this paper are
that (a) partial information cannot correct the market and (b) that the issue is not the
level of information about the whole contract, but each term should be examined
separately.
The intuition behind the argument is as follows. An inefficient term is a contract
term which reduces the cost to the seller by creating a smaller reduction of the value
of the contract to the consumers. When such a term is included in the contract, the
seller fully internalizes the cost reduction. However, the proportion of informed
consumers prorates the reduction in demand. For example, a term that reduces the
value of the contract to the consumers by €10, reduces the highest price informed
consumers are willing to pay (the reservation price) by €10, while leaving the
demand of the uninformed unchanged. If only half of the consumers are informed,
the term would reduce the average reservation price by only €5. In other words, the
seller internalizes only half of the value’s reduction. By introducing the term, and
simultaneously reducing the price by €5, she can assure that her sales stay in the
same level. Thus, she will include the term as long as it saves her more than €5 in
costs, even if the term reduces the contract’s value to the consumers by €10. Now, in
order to maximize her profits, the seller adjusts the level of one-sidedness of each
term to the proportion of consumers who are informed about the term.
Obviously, if 90% of the consumers are informed, introducing such a €10 term is
worthwhile only if the term reduces the seller’s costs by more than €9. In some
cases this is not the case. But in most cases the seller can adjust the content of the
term to meet this criterion. And if adjusting one term is more difficult, other terms
might be available. Thus, there is no threshold beyond which the portion of
informed consumers assures full internalization of terms’ costs and benefits. A
gradual increase in the portion of informed consumers will gradually increase the
internalization of the cost of inefficient terms, while full information result can only
be achieved if full information actually exists.
This does not mean that contracts’ quality deteriorates indefinitely. Many
stipulations come into force in every contractual relationship and thus are known to
all but a few first-time consumers. For example, if instead of offering a door-to-door
service, a delivery contract stipulates that the recipient has to collect the shipment
from an office of the carrier, this stipulation will be quickly known to every user of
the service. Hence, the seller (here, service provider) would only insert such a
4 Of course, terms might also be inefficient if they increase the cost to the seller more than they benefit
consumers. However, no rational seller would insert such terms to the contract, even if all the consumers
were fully informed. If a term costs the seller €5 but increases the value to consumers by only €4, the
seller could always profit by removing the term and reducing price by €4. Such an action will assure her
that the quantity demanded remains the same while the profits for each contract increases by €1. Thus,
when sellers set the terms there is no reason to expect pro-consumer inefficient terms.
Eur J Law Econ (2007) 24:119–136 121
123
stipulation if consumers were better off with lower prices than they are with the
door-to-door service.
In addition, most courts refuse to enforce terms which clearly lack economic
sense and could only be explained as an exploitation of information deficits. If a
carrier stipulated in the fine print that consumers must compensate her and third
parties in case of an accident, even if the accident was the carrier’s fault, the term
would not be enforced. Obviously, the carrier is a better cost avoider in such a case,
and most courts would see things that way.5
As a result, the seller is only able to rely on consumers’ information deficits
regarding terms which can be viewed as relevant to the contract but may still be
unknown to many consumers. The classic example is terms that transfer contractual
risks to consumers. These terms come into force only rarely, and thus are unknown
to many consumers. Often the carrier can show an economic rationale for such a
term, since most risks are affected by the behavior of both parties, and it is hard to
prove who the best risk bearer is. This paper shows that in the absence of judicial
review, the supplier will adjust the risk allocation of these terms inefficiently.6
The rest of the paper is organized as follows: Part II presents a model of a
monopoly which offers the same standard form contracts to both informed and
uninformed consumers. This part shows that in the absence of full information, the
seller offers suboptimal contract terms. Part III examines the incentives of
consumers to read the contract. A few concluding remarks appear in Part IV.
2 The model
A non-discriminating, rational, profit-maximizing monopoly (the seller) supplies a
service subject to terms of a nonnegotiable standard form contract.7 The price of the
service (p) and the basic features of the contract are known to all. The personal
preferences of the basic features of the service, W, is distributed in the population by
a distribution function F(W), i.e., the portion of the population that estimates the
value of the feature as at most W is given by F(W).The seller has to decide whether to include in the contract a term that will only be
observed by the reading consumers. The seller can construct the term in different
ways, and by doing so she determines the effect of the term on the value of the
5 On the other hand, courts might enforce a term which limits the carrier liability for damages to the
shipment, since covering for such damages would require senders of inexpensive materials to subsidize
the insurance for those who send expensive material. Requiring such cross subsidy might result in adverse
selection, and hence it is less clear whether the inclusion of such a term is the result of consumers’
information deficit.6 In some cases, informed consumers might angrily reduce their willingness to pay beyond the direct cost
of the inefficient term, and thus counterbalance the ignorance of the uninformed. When such anger is
sufficiently influential it might help to discipline the seller.7 Throughout the article I assume that sellers know and understand the terms of the contract they offer,
which is usually the case since, unlike consumers, sellers draft the standard contract. Moreover, sellers
draft the contracts with care, knowing that the form will serve them for numerous transactions. Random
mistakes might, of course, occur even from the seller’s side, but the problem discussed here refers to
systematic lack of information which one side (the seller) can rely on.
122 Eur J Law Econ (2007) 24:119–136
123
service to the consumers. For example, the seller can add an exemption clause for
damages caused by delay in supply and choose the length of the exempt delay and
the circumstances. Obviously, the longer the exempt delay and the broader the
exempt circumstances are, the lower the value of the contract to the consumers is.
Yet limited exemption clauses might be beneficial to both parties. A term exempting
the seller from losses caused by short delays in supply that were not a result of the
seller’s negligence can save expected costs of future litigation, and thus reduce
the cost by more than the reduction in value. On the other hand, a term exempting
the seller from long delays that she causes deliberately prevents internalization
of the damage she causes, and hence is inefficient. The seller can control the scope
of the exemption, and thus control its effect on the consumers.
Let k denote the quality of the contract in monetary terms, i.e., the effect of the term
on the contract’s value to the consumers. For example, the quality of a term that
reduces the value of the contract to the consumers by €10 is k = – €10. Since we are
interested in quality and not in diversity, k is assumed to be the same to all consumers.8
Thus, the reservation price of each informed consumer is a sum of his personal
preference for such a service when the hidden term is not included, W, plus the benefit
the consumer gains from the quality, k. In other words, k represents the relative value of
the term to consumers, compared to the default rule which would apply in the absence
of the term. Since our concern is about terms that harm consumers, k in our analysis is
usually negative (when the term is not included at all k = 0). However, when the
default rule is inefficiently pro-seller, even a pro-consumer stipulation (k [ 0) might
be inefficient, if it does not deviate sufficiently from that rule.
The seller’s marginal cost, c(k) is constant for quantity and it depends on the
quality of the contract k. Thus, the total cost of the seller is Q � c(k) + C0, when Q is
the quantity of services she sells, and C0 is the fixed cost. From now on, for
simplicity, I assume that there are no fixed costs, i.e., C0 = 0. This assumption does
not affect the result of the model. The seller’s revenue depends on the quantity Q,
and the price P, that is P � Q. The marginal cost function c(k) is increasing and
convex for quality,9 i.e., c0(k) [ 0 and c00(k) [ 0. The only source of information
about the content of the hidden term is reading. At any given time, every consumer
buys at most one service.
2.1 The quality when all consumers are fully informed
Let us assume first that all consumers are fully informed of the quality k. Let F(P)denote the portion of the population that estimates the price of the contract to be too
8 When k is different for different consumers, then a change in k might reduce the welfare in some
transactions but increase it in others. For simplicity, I do not refer here to the possibility that different
consumers have different preferences regarding the hidden terms.9 The convexity of the cost function can be explained using the example of the exemption clause. When
the quality of the contract is low, because of a broad exemption clause, the seller can, at a relatively small
cost, improve the contract substantially. By narrowing the scope of the clause not to include substantial
risks to the consumers that can be cheaply prevented (e.g. holding minimal stocks to reduce the risk of
long delays in supply), the seller can increase the quality considerably at low costs. On the other hand,
when the quality is already high, additional increase would impose more substantial expenditure.
Eur J Law Econ (2007) 24:119–136 123
123
high when the quality is k = 0. Thus the proportion of potential consumers that
purchase the contract of quality k = 0 is Q = 1 – F(P). If we normalize the maximum
possible quantity purchased by the population to be 1, then this is the demand curve.
By generalizing the demand curve for every k when all consumers are fully
informed, we get:
Q ¼ 1� FðP� kÞ ð1aÞ
In this equation, P is the price the seller can charge if she wants to sell Q units of
quality k when all consumers are fully informed.
We can write (1a) in an inverted form, when P is the dependent variable:
P ¼ F�1ð1� QÞ þ k ð1bÞ
When F–1(�) is the inverted function of F(�).The monopolist’s profits are:
P ¼ [F�1ð1� QÞ þ k]Q� Q � c(k) ð2Þ
Where P is the profit of the monopolist, [F–1(1 – Q) + k]Q is her revenue and
Q � c(k) is her cost. For a given Q, the seller can maximize her profits when
oP=ok ¼ 0, and o2P=ok2\010. The second order condition requires c00(k) [ 0,
which is true by assumption for every k, and the first order condition is:
c0ðkÞ ¼ 1 ð3Þ
That means that no matter what Q will be chosen by the seller, it is in all cases
worthwhile for her to provide the same quality k. This is also the quality level that
maximizes the overall welfare. If k is lower, and thus c0(k) \ 1, the seller can increase the
quality and price by €1 without affecting the quantity demanded, while increasing her
costs by less than €1. If k is higher, and hence c0(k) [ 1, the seller can reduce the quality,
and by so doing save costs that are higher than the reduction in quality. Hence, when all
consumers are fully informed of k, a monopolistic seller creates no quality distortions.11
The intuition behind this result is plain. If all consumers know the quality, they
will adapt their willingness to pay accordingly. Thus, the seller would adjust the
quality to the level in which the marginal cost of quality equals the marginal benefit
of the quality for the consumers, which is by definition equal to €1.
2.2 The quality when some consumers are uninformed
Assume now that only a proportion of the population a (0 £ a £ 1) is informed of
the content of the hidden term. In all other respects, including the distribution of the
10 To be more precise, when maximizing her profits, the seller forces oP=oQ ¼ 0 and o2P=oQ2\0.
These constraints help him determine which quantity Q it is optimal to produce. With respect to this
optimal quantity, an optimal quality k is determined. These additional constraints on the quantity do not
affect the result regarding the derivative of the cost function c(k).11 For a similar result in a competitive market with fully informed consumers, see Cooper and Ross
(1984)
124 Eur J Law Econ (2007) 24:119–136
123
reservation price, the uninformed are a represented subgroup of all consumers. In
other words, there is no correlation between the reservation price of a consumer and
the probability that this consumer is informed.12
Let QI denote the quantity demanded by the informed consumers, and QU by the
uninformed. Thus, the overall demanded quantity is the sum of the quantity
demanded by both groups, i.e.,13
Q ¼ Q1 þ QU ð4Þ
For simplicity, assume that the demand is linear.
Assuming linearity, the demand of the informed consumers is
QI ¼ a � A� ðP� kÞb
ð5aÞ
or
P ¼ A� b
a� QI þ k ð5bÞ
with A and b determining the intercept with the axis and the slope of the function.
The uninformed have similar preferences but they do not know the real k, hence
they decide according to their estimation of the quality. Let kU denote this
estimation, and assume that this estimation is the same for all consumers (an
assumption which is relaxed later). For the time being, the exact value of ku is
unimportant, as long as the seller cannot affect it. The demand of the uninformed is
hence:
QU ¼ ð1� aÞ � A� ðP� kUÞb
ð6aÞ
or
P ¼ A� b
1� a� QU þ kU ð6bÞ
Using (4), (5a) and (6a) it can be shown that
QU ¼ ð1� aÞ � Q� ð1� aÞ � ab
� k � kUð Þ ð7Þ
For the seller, interested in maximizing her profits, QU can thus be determined by
her choice of Q and k.
12 This assumption is often used in the literature. See for example Wilde and Schwartz (1979b), Cooper
and Ross (1985).13 In order to avoid problems of range definitions and to simplify the analysis, I assume that the results do
not occur near the upper edge of the demand curve, when the monopolist might sell only to the informed
consumers or only to the uninformed ones.
Eur J Law Econ (2007) 24:119–136 125
123
The seller’s profit is
YðQ; kÞ ¼ PðQ; kÞ � Q� c(k) � Q ð8aÞ
or
YðQ; kÞ ¼ P QUðQ; kÞð Þ � Q� c(k) � Q ð8bÞ
Using (6b) and (7) it can be shown that 8b is maximized when
c’(k) ¼ a ð9Þ
This result means that for every given Q, a monopolist supplier will adjust the
quality of the hidden term to the point where an additional reduction of €1 in her
cost reduces the quality to the consumers by €a.14 As a consequence, only when allconsumers are informed (a = 1), will the monopolist form an efficient contract with
c0(k) = 1.
The intuition behind this result is as follows: Each €1 reduction in the value of
the service reduces the demand of the informed by €1, while not reducing the
demand of the uninformed at all. If half of the consumers are uninformed, the
aggregated demand will be reduced by €0.50. Thus, as long as the seller can reduce
her cost by €1 by reducing the quality by less than €2, she would do so.15
3 The decision to read the contract
In this part I examine the factors that might affect consumers’ decision to read the
contract in order to know the quality of its terms. Consumers with very high
willingness to pay are likely to refrain from reading the contract. These consumers
know that even if the quality of the contract, k, is much lower than the quality they
expect kU, they are still likely to buy, and thus there is no point in the effort of
reading. On the other hand, when the reservation price adapted by kU is closer to the
price set by the seller, the chances that the quality will affect the decision to buy are
14 If we relax the linearity assumption, the result of (9) will be only slightly different. If the demand
curve is convex, and k \ kU (k [ kU) then c’(k) [ a c’(k) \ að Þ. That means that for a convex demand
curve, the informed consumers influence the result by more than their portion a when the uninformed
consumers over-estimate the quality, while their influence will be less than their portion a when the
uninformed consumers under-estimate the quality. For a concave demand curve, the result is vice versa.
However, since the convexity is not likely to be significant in the small range affected by the term, and
since the difference between the real quality and the estimated quality would not be significant either, the
result would not vary substantially from a.15 Even when the seller cannot adjust the quality gradually, he might find inefficient terms to be
beneficial. In the above case, the seller would benefit from a term that reduces the value of the contract by
€100 as long as this term saves him more than €50 (say, €51). One can see that the profit of the seller
increases after this process if the seller, simultaneously, reduces the price by €50. In this case, the quantity
demanded would be the same as before, but for each product sold, the seller will profit an additional Euro,
because the price is down by €50, while the cost is down by €51. Note that the seller might still adjust the
quantity and the price further for maximal exploitation of the term, but for the purpose of this article it is
enough to show that such a term would increase profits by at least €1 per unit.
126 Eur J Law Econ (2007) 24:119–136
123
higher. Thus, in such cases, consumers are more likely to read the contract before
signing. In this part I present the main factors affecting the decision to read.
3.1 The factors affecting the decision to read
Let us observe one consumer. Define this consumer’s willingness to pay for a
contract with a quality of k = 0, as W. If the consumer is fully informed, he will buy
the contract when P £ k + W. However, assume now he is not informed and he can
choose whether to pay the costs of reading and understanding the contract or not.
The costs of reading the contract are r, and by bearing them he will know the real
quality of the contract k. If he does not read the contract, he will have to rely on his
uninformed estimation of the quality of the contract kU. Assume that the consumer
believes that the real quality of the contract is somewhere between kU – z and
kU + z, while in this range every number is equally possible.16 Further assume risk
neutrality, thus the consumer just maximizes expected benefits.
Let us define G as the expected net gain for the consumer from buying the service
without reading. Thus
G � W þ kU � P ð10Þ
If G ‡ z, the consumer will not read and will buy the service, because according
to his estimation the contract will be beneficial in any event. Even if the contract’s
quality was in the lowest place in the estimated zone, i.e., if k = kU – z, he would be
better off buying the contract, according to his beliefs.17
A more complicated situation arises when W + kU – z \ P £ W + kU (i.e., when
0 £ G \ z). In this case, though the expected gains from the contract are positive,
the contract may also be a bad deal. The consumer then has to decide whether
checking the contract is worth the costs.
If the consumer reads the contract, the probability that he will buy is (G + z)/2z.
This can be seen in the above figure. If after reading, the reservation price
(W + k) is between the points P and W + kU + z then the consumer will sign the
contract. The distance between these points is G + z. The portion of this distance out
of the whole possible range is (G + z)/2z.
16 It is more likely to assume a normal distribution of kU around k. However, this would extremely
complicate the calculation without adding a thing to the substance of the result.17 For consistency, I assume that when the consumer is indifferent between buying or not, he buys the
service; if he is indifferent between reading and not reading, he does not read.
Eur J Law Econ (2007) 24:119–136 127
123
His expected gain if he buys after reading is (G + z)/2, which is the average of gains
from all contracts which are worthwhile for him between P and W + kU + z. This average
is higher than G because after reading, there is no chance that the consumer will enter a
contract which is not worthwhile, i.e., a contract whose worth is less than P. His expected
net gains from deciding to read is equal to the probability he will buy after reading, times
the expected benefits from buying after reading, minus the costs of reading. Hence
Ur ¼Gþ z
2z� Gþ z
2� r ¼ ðGþ zÞ2
4z� r ð11Þ
when Ur is the expected gain to the consumer when he reads the contract. The
consumer will read the contract if, and only if, the Ur is higher than the expected
gains from signing without reading G; meaning if
ðGþ zÞ2
4z� r [ G ð12Þ
If W + kU + z ‡ P [ W + kU (i.e. z £ G \ 0) the consumer will not buy unless
he decides to read and after reading he finds out that the contract is worth the price.
The consumer will read if the expected net gains from reading are positive. The
expected gains from reading are the same as in (12), only that this time it should be
compared to 0 and not to G (because the consumer will not buy if he does not read
when G \ 0). Thus the consumer will read if:
ðGþ zÞ2
4z� r [ 0 ð13Þ
Of course, if P \ W + kU – z, (i.e., G \ –z) the consumer will not read and will
not buy.
We can now summarize the conditions for which the consumer will read the
contract.
The consumers will read the contract if, and only if:
– z \ G \ z and
ðGþ zÞ2
4z� r [
G if G [ 0
0 if G� 0
(
Solving (11) and (12) for G gives us:
�zþ 2ffiffiffiffirzp
\G\z� 2ffiffiffiffirzp
ð14Þ
If this condition is fulfilled, and only then, the consumer will read the contract.
3.2 The readers ability to correct the market failure
The result in (14) helps us estimate whether and when the reading consumers can
serve to correct quality distortion in the market. From this result we can deduce the
following conclusions:
128 Eur J Law Econ (2007) 24:119–136
123
3.2.1 Conclusion 1
If r\1=4z, and for all consumers in the margins, (i.e., consumers who would change
their choice if they are informed of a change in quality) kU is not different from k by
more than z� 2ffiffiffiffirzp
, then every small change in the quality k will be fully
internalized. This is because all the consumers who may change their decision due
to a change in quality will read the contract. As shown in the first part, when this
happens, the supplier has incentives to offer an efficient contract. This shows that
under certain conditions a monopolist has the right incentives to produce the
efficient quality.
However, from the optimistic outcome in conclusion 1, we can also find the
pessimistic and more realistic outcome that occurs when the conditions for
conclusion 1 are not fulfilled:
3.2.2 Conclusion 2
When r [ 1=4zthen G does not fulfill the requirements of condition (14). If r [ 1=4zfor all consumers, no consumer would read the contract, no matter how close G is to
zero. In such a case, no knowledge in the market about the contract terms will exist.
If some knowledge is available from other sources—like word of mouth
advertisement or repeated purchases—then this knowledge will spread randomly
and will not necessarily reach the marginal consumers, an issue that is discussed in
the next part. Thus when r [ 1=4z, the model presented in the previous paragraph
better represents the expected outcome.18
3.2.3 Conclusion 3
If the different consumers have different estimations as to the size of k, then again
the marginal consumer will not play their market-correction role. The intuition for
this result, whose formal proof is omitted, is shown in the following example. Let us
assume a service which costs P = 100, given at a real quality of k = 0. Assume that
the potential consumers are divided into three main groups, A, B, and C, according
to their willingness to pay for such a contract, when they are fully informed. The
consumers for whom the contract is worth between 85 and 95 are in group A. For
consumers in groups B and C, the product is worth between 95 and 105, and
between 105 and 115 respectively. Each of these three groups is divided into three
equal sub-groups, according to their estimation of the quality of the contract. The
second row shows the consumers that rightfully estimate the contract’s quality to be
0. The consumers in the first row underestimate k to be –10; those in the third row
overestimate it to be +10. Thus, there are nine equal subgroups of consumers
according to their willingness to pay and their estimation of the contract’s quality.
18 For a graphical example of conclusions 1 and 2, when different consumers have different costs of
reading, see the Appendix.
Eur J Law Econ (2007) 24:119–136 129
123
WA = 90 ± 5, WB = 100 ± 5 WC = 110 ± 5
kU1¼ �10 WA + kU1
= 80 ± 5 WB + kU1= 90 ± 5 WC + kU1
= 100 ± 5
kU2¼ 0 WA + kU2
= 90 ± 5 WB + kU2= 100 ± 5 WC + kU2
= 110 ± 5
kU3¼ 10 WA + kU3
= 100 ± 5 WB + kU3= 110 ± 5 WC + kU3
= 120 ± 5
When z� 2ffiffiffiffirzp ¼ 5, the consumer will read the contract if, and only if, his
expected gain from it without reading is lower than 5 and higher than –5
(–5 \ G \ 5). Since the contract’s price is 100, the consumers who will read the
contract are those who estimate the contract as worthwhile between 95 and 105 for
them. One can easily see that these consumers are a third of each group: subgroup 3
of group A, subgroup 2 of group B and subgroup 1 of group C (marked in gray in
the table). In this example, due to the mistakes in the estimation of the quality, the
informed consumers are a random portion of all consumers and not only the
marginal consumers (in this example the marginal consumers are the consumers in
group B). This again brings us back to apply the model that was developed earlier.
The informed consumers are one third of the consumers randomly distributed
among the consumers.19
3.2.4 Conclusion 4
When a big reduction in k is possible, the marginal consumers cannot correct the
market failure. If the seller can alter the quality k and price P without reducing her
profits per unit, in a way that the new price would cause all the previous readers to
buy the service without reading, it is always worthwhile for her to do so. In order to
see it, let us look again at the table shown above, but this time I assume that all
consumers estimate correctly the product quality. Thus, the three groups exist as in
the table below. In this case all consumers in group C will buy without reading. The
consumers in group B will read the contract and half of them will buy the service.
Consumers in group A will not read and will not buy.
WA = 90 ± 5 WB = 100 ± 5 WC = 110 ± 5
kU2= 0 WA + kU2
= 90 ± 5 WB + kU2= 100 ± 5 WB + kU2
= 110 ± 5
Now assume that the seller can reduce the quality k by 20, and by doing so she
reduces the cost per unit by 10. For simplicity, assume that all the change in costs is
reflected in the price; thus the price of the service is now only P = 90, and the
quality is k = –20. The consumers cannot detect the change in k, since they do not
know, by assumption, the supplier’s cost function. However, they react to the
19 See a graphical example, explaining this conclusion for a broader range of differences in estimations,
in the Appendix.
130 Eur J Law Econ (2007) 24:119–136
123
change in P. Now all consumers in group B, who still estimate k to be kU = 0, will
buy without reading. Consumers in group A will read, and after doing so none of
them will buy. In group C, no change will occur and all the consumers of that group
will buy without reading. Thus, half of group B who did not buy under the previous
set of price and quality will buy under this inferior set. The seller inserted an unfair
clause to the contract, and by doing so she increased her surplus by 33.3%, though
all consumers have correctly estimated the quality before the change. It can be
easily shown that the same result will occur if the consumers overestimate the
quality before the change and correctly estimate it afterwards.20 It is always true that
if the seller can reduce her costs, and thus her price by more than z� 2ffiffiffiffirzp
, then it is
worthwhile for him to do so, no matter how big the corresponding reduction in
quality should be.21
In summary, it seems that the more complicated assumption, according to which
the informed consumers are not a random portion of the consumers, but rather are
those who find it worthwhile to read the contract, does not add a lot to the model.
Thus, the conclusion which was drawn in the previous section is still valid.
Knowledge about the quality that is in the hands of only some consumers cannot
create an efficient result.
4 Concluding remarks
4.1 Different potential terms and reputation
Sellers have many available terms at their disposals. For each term, a different
proportion of consumers would be informed. Terms that become known to every
consumer after the transaction, will be known to many potential consumers even
without reading the contract. Terms that take effect only rarely (like exemption
clauses) or terms whose effect is hidden from the consumers (like bank
commissions which are not presented clearly in the bank’s statement), will usually
be known to a smaller proportion of consumers. The seller has, thus, two tools in her
hands. She can choose the type of term, and prefer terms that become known only
rarely, or she can choose the exact content of the term, adjusting it to the effect on
the costs and quality, considering the level of knowledge for such a term. Hence,
sellers can always construct profitable inefficient terms.
20 Many models of asymmetric information assume that consumers know the seller’s cost function and
his alternatives and thus can infer that such quality-reducing terms will be introduced. If this is the case,
then consumers will infer that the inferior term is included in the contract, and thus estimate the quality of
the contract correctly after inclusion. This model uses a less strict assumption about consumers’
knowledge, and shows that the quality reduction would still occur whether the consumers correctly
estimated the final quality of the contract or not. For more details see section IV B.21 This conclusion is correct if there is no knowledge that is randomly transferred to a portion of the
consumers, namely, that the only informed consumers are those who find it worthwhile to read the
contract according to their estimation of its quality. If some information exists in the hands of a random
portion of the consumers, in addition to the knowledge of the readers, the previous model can be applied
for a big change in k.
Eur J Law Econ (2007) 24:119–136 131
123
Reputation can increase the proportion of informed consumers. Some consumers
might learn about the quality of contract from word of mouth, newspaper stories and
similar sources which generate the seller’s reputation, and thus be informed of the
quality without reading. In few industries such mechanisms might have a substantial
effect. For example, in the car industry, the quality discoveries are quickly revealed
to a very big proportion of the consumers through private information companies,
especially car magazines and motor-clubs. Yet in most cases at least some terms can
be hidden for a substantial period of time having a very limited effect on the demand.
As Shapiro (1982) showed, even if the term is ultimately made known to everyone
(and this is rarely the case), the time lag assures that the seller would behave as if
information is partial. From time to time a term can be made public because of a
successful campaign of a displeased consumer. But even in these rare cases in the life
of a seller, the seller can publish the changes in policies or offer alternative contracts
as well after the damaging publication, substituting one harming term with another.
4.2 The estimated quality by the uninformed
Usually asymmetric information models assume that uninformed consumers know
enough about the cost of the seller and the market structure, and thus, at equilibrium,
correctly deduct the quality of the product (see, for example, Akerlof 1973). Such
assumptions seem unrealistic. Consumers with no knowledge of the content of the
contract they sign are unlikely to know so much about the costs of the seller and the
demand she faces. In any case, the case where kU is equal in equilibrium to k is
simply a special case of the more general model presented above. Hence even if we
assume that such knowledge exists, the result of the model remains the same.
That does not mean that uninformed consumers know nothing about the sellers’
interests. From experience, they know that standard form contracts are often one-
sided. They might even know that sellers are incentivized to reduce the contract
quality in such a contract, even if they cannot know exactly how. Thus, they might
be less willing to buy a service with such a standard form contract. In other words,
the demand for the service might be lower because of the information asymmetry,
and the consumers’ belief that the hidden terms are unfavorable. Hence, sellers
might also suffer from the information deficit. The effect of the information
asymmetry on sellers is indeterminate, because in some cases consumers might
underestimate the means of reducing the contract’s quality. In such cases, the
demand is not sufficiently affected by consumers’ concerns, and the seller’s gain
from the terms exceeds her losses from the decrease in demand. In other cases,
though, consumers’ concerns might more accurately reflect reality, or even be too
pessimistic. In these cases, the demand is substantially affected by consumers’
concerns. Sellers in these markets would also lose from the information deficit.
Courts can thus play a positive role in correcting the market. If courts can detect and
invalidate inefficient pro-seller terms, both parties can gain. Sellers will have fewer
options to reduce quality inefficiently and consumers will learn to trust such contracts,
and thus be more willing to pay for them. Obviously, if courts often make mistakes,
such a policy might be counterproductive. Thus, one cannot infer that encouraging
132 Eur J Law Econ (2007) 24:119–136
123
courts’ intervention is always the solution. Yet the prevailing view which relies on
informed consumers to do the job is also too optimistic, and thus cannot, in itself,
justify the calls for judicial restraint in reviewing standard form contracts.
4.3 More competitive markets
Though the model refers to a monopoly, most of the analysis is relevant to more
competitive markets as well. In most markets, sellers have some market power, due
to differences in the product or terms. Often they create market power by non-price
competition. Thus, many of the factors affecting a monopoly are likely to affect a
seller in monopolistic competitions. It is true that in monopolistic competition,
reduction in quality would drive many informed consumers to competitors. Yet the
price decrease that the quality reduction allows would drive many uninformed
consumers from competitors to the seller. A more thorough examination of
competitive markets should be addressed separately, but for now it is enough to say
that competition, in itself, does not necessarily alter the conclusion.
Acknowledgements This paper greatly benefited from comments and criticisms of Omri Ben Shahar,Gerrit De-Geest, Hans-Bernd Schaefer, participants of The Annual Conference of the Israeli Law andEconomics Association, Tel Aviv, 2003 and the participants of the Annual Conference of the EuropeanAssociation of Law and Economics, Nancy 2003. Ziv Glazberg provided excellent research assistance.
Appendix-graphical examples
The demand when some consumers are uninformed
Note: In the following graphs, the index 1 (D1, MC1, MR1) refers to the situation that
exists if the seller does not include the quality reducing term in the contract; and index
2 (D2, MC2, MR2) refers to the final equilibrium after the term is included. Variables
that do not change due to the change in quality are marked with no index (D, MC, MR).
In the above example, the portion of informed consumers is 2/3. A reduction in
quality of €3 causes a reduction of €3 in the willingness to pay by the informed
consumers (see Dk in the second graph) but no reduction in the willingness to pay by
the uninformed consumers. Therefore, the aggregate demand curve is €2 lower than
before the reduction in quality. If, by reducing quality by €3, the seller can reduce her
Eur J Law Econ (2007) 24:119–136 133
123
cost by more than €2, she can sell the same quality as before, for €2 less, and increase
profits. (See this result in ‘‘The choice of the monopoly, when only a portion of
consumers is informed’’).
The choice of the monopoly, when only a portion of consumers is informed
In this example, the two aggregated demand curves (D1 and D2) are presented as in
the previous page. MC1 and MC2 are the marginal cost curves before inserting the
unfair term and after, respectively. MR1, and MR2 are the corresponding marginal
revenue curves. The upper rectangle, printed in a bolded line, is the profits before
inserting the unfair term, and the lower one, which is filed in with dots, is the profits
after. Since the cost saving of inserting the term is smaller than a � Dj (MC1 –
MC2 = 1 \ 2 = a � Dj) the monopoly would loose from inserting the term. That
can be seen from the difference in size of the profit’s rectangles.
This example is similar to the one on the left, only here the reduction in cost after
inserting the unfair term is of €2.5 per contract (i.e. MC1 – MC2 = 4 – 1.5 = 2.5).
Since this reduction in cost is bigger than the reduction in quality times the portion of
informed consumers (2.5 [ a � Dj, since a � Dj = 2) it is profitable to insert the
unfair term, although it create more losses to consumers than benefits to the supplier.
Graphically it can be seen from the size of the rectangles. The upper rectangle,
representing the profits before inserting the term is bigger than the lower one.
134 Eur J Law Econ (2007) 24:119–136
123
Different estimations of the quality
On this page, the uninformed consumers have different estimations of the quality of the
contract. Ten percent of the consumers underestimate the quality by€1 (left graph), 20%
correctly estimate it (second from the left), and another 20% overestimate it by€3. Since
there is no correlation between the estimation of quality and the willingness to pay for the
main properties of the contract, the demand of each sub-group is still linear.
The aggregated demand curve of the three sub-groups of uninformed consumers
(the graph on the right) is similar to the demand curve these consumers would have
if each of them overestimated the quality by €1 (if we disregard the upper edge of
the curve, above the breaking point). This is because, on the average, they
overestimate the quality by €1. Now we can introduce the demand of the remaining
50% of the consumers—the informed consumers. This is done in the next page.
The demand curve of the informed consumers is presented in the graph on the left.
In the initial state, this curve is D1. If the seller reduces the quality by €3, their
willingness to pay will be reduced by €3, and their demand will be D2. The aggregated
demand curves, on the right-hand side shows that a reduction in quality of €3 caused
the demand curve to be €1.5 lower, as anticipated by the model when 50% of the
consumers are informed.
Two new features were presented on these two pages. First, here the uninformed
consumers have different estimations of the contract’s quality. Second, the average
estimation is wrong (here, overestimation of €1 on the average at the initial point).
Yet the result stays unchanged. The monopoly would insert an unfair term if, and
Eur J Law Econ (2007) 24:119–136 135
123
only if, the cost-reduction of doing so is higher than the reduction in quality, times
the portion of informed consumers (i.e., Dk � a \ Dk). As long as there is no
correlation between the willingness to pay for the main properties and the
estimation of quality, different quality estimation does not change this result.
References
Ahdieh, R. B. (2005). The strategy of boilerplate. Emory Law and Economics Research Paper No. 05–22.
Akerlof, G. A. (1973). The market for ‘lemons’: Quality uncertainty and the market mechanism. TheQuarterly Journal of Economics, 84(3), 488–500.
Cooper, R., & Ross, T. W. (1984). Prices, product qualities and asymmetric information: The competitive
case. The Review of Economic Studies, 51(2), 197–207.
Cooper, R., & Ross, T. W. (1985). Monopoly provision of product quality with uninformed buyers.
International Journal of Industrial Organization, 3, 439–449.
Cooter, R., & Ulen, T. (1997). Law and Economics, 2nd edn. Reading: Addison-Wesley Longman.
Cruz, R. T., & Hinck, J. J. (1996). Not my brother’s keeper: The inability of an informed minority to
correct for imperfect information. Hastings Law Journal, 47, 635–676.
De Geest, G. (1994). Economiche Analyse van het Contracten–en Quasi-Contractenrecht: Een onderzoeknaar de wetenschappelijke waarde van de rechtseconomie. Maklu: Antwerpen.
De Geest, G. (2002). The signing-without-reading problem: An analysis of the European directive on
unfair contract terms. In Hans-Bernd Schafer & Hans-Jurgen Lwowski (Eds.), Konsequenzenwirstschaftsrechtlicher Normen (p. 213).
Gillette, C. P. (2004). Rolling contracts as an agency problem. Wisconsin Law Review, 2004, 679–722.
Gilo, D., & Porat, A. (2006). The hidden roles of boilerplate and standard form contracts: Strategic
imposition of transaction costs, segmentation of consumers and anticompetitive effects. MichiganLaw Review, 104, 983–1031.
Goldberg, V. P. (1974). Institutional change and the quasi-invisible hand. Journal of Law and Economics,17(2), 461–492.
Hillman, R. A., & Rachlinski, J. J. (2002). Standard-form contracting in the electronic age. New YorkUniversity Law Review, 77, 429–495.
Kornhauser, L. A. (1976). Unconscionability in standard forms. California Law Review, 64, 1151–1183.
Korobkin, R. (2003). Bounded rationality, standard form contracts, and unconscionability. University ofChicago Law Review, 70, 1203–1295.
Posner, R. A. (1992). Economic analysis of law, 4th edn. Boston: Little, Brown and Company.
Schwartz, A., & Wilde, L. L. (1979a). Intervening in markets on the basis of imperfect information: A
legal and economic analysis. University of Pennsylvania Law Review, 127, 630–682.
Schwartz, A., & Wilde, L. L. (1979b). Equilibrium comparison shopping. The Review of EconomicStudies, 46, 543–553.
Shapiro, C. (1982). Consumer information, product quality, and seller reputation. The Bell Journal ofEconomics and Management Science, 13(1), 20–35.
Trebilcock, M. J., & Dewees, D. N. (1981). Judicial control of standard form contract. In Paul Burrows &
Cento G. Veljanovski (Eds.), The economic approach to law (pp. 93–119). London: Butterworth.
136 Eur J Law Econ (2007) 24:119–136
123